Background/Purpose: Whilst the exact aetiology of rheumatoid arthritis (RA) remains unclear, current concepts suggest an environmental link with possible microbial triggers.  Studying individuals at-risk of RA allows a unique opportunity to explore possible relationships.  Furthermore, the identification of predictors of progression from at-risk to RA is of great clinical utility.   Triggering receptor expressed on myeloid cells (TREM) modulates the innate immune response by either amplifying or dampening the toll like receptors (TLR) induced signalling¹.  Given the upstream nature of soluble-TREM (sTREM), and its close proximity to TLR in innate immune response to microbial triggers, we hypothesis that sTREM could be one of the first indicators of progression to RA.   

Methods: sTREM was measured using Human TREM-1 Quantikine ELISA (R&D Systems, Minneapolis) in 32 at-risk individuals; 16 of whom progressed (ACPA+ Prog) and 16 whom did not (ACPA+ NProg), and 16 early RA (ERA) patients.  The ACPA+ cohort had no evidence of ultrasound synovitis at baseline. Binary logistic regression models were conducted on at-risk patients. Model discrimination was assessed using classification plots and the area under the ROC curve. An optimum cut-off for the test variable, plus associated sensitivity and specificity achieved, were also determined. Further analyses were conducted on the same sample augmented with ERA patients, to assess the correlation between TREM, CRP and CCP titre values in this group; and on ERA patients only, to assess the correlation between TREM and DAS28 values in this group.  

Results: Duration of follow-up was 0 to 133 months, mean duration of 36 months (SD 31.2).  Mean time to RA diagnosis in those that progressed was 11.5 months (SD 14.3).   

A binary logistic regression model found that TREM was significantly associated with progression (p=0.009). The OR of 1.006 (95% CI 1.002 to 1.011) indicated that at best estimate, the odds of progression increased by 0.6% with each increase of 1 unit of TREM.  The model classified 80.6% of cases correctly, with 12 out of 16 (75.0%) of ACPA+ P cases and 13 out of 15 (86.7%) of ACPA+ NP cases correctly classified. The area under the ROC curve was 0.825 (95% CI 0.674 to 0.976).  Optimum combinations of sensitivity (81.3%) and specificity (80.0%) were obtained from a cut-off TREM value of 521.3 units.  There were no substantive correlations between TREM, CRP values and CCP values.

Conclusion: In this first assessment high level of sTREM is significantly associated with disease progression in at-risk individuals.  Furthermore, higher sTREM values are associated with greater odds of progression. It has potential to offer insights regarding disease pathogenesis and warrants further research.  References: 1. Bouchon A et al, Nature 2001